Infinite Context LLMs: Going Beyond RAG with Extended Minds • Hacker News Board

Re: Infinite Context LLMs: Going Beyond RAG with Extended Minds

#41

Post by bsenftner » Wed, Nov 15, 2023, 2:56 PM UTC

mike_hearn wrote: post #40

bsenftner wrote: post #35

Earlier quoted context omitted.

It's open source with a noncommercial use license.

Huh. I wonder if that meets the definition of open source. I didn't notice the special license.

I additionally noticed an Apache license on their models at Huggingface, but the blog post states the work described by that blog post is open source for noncommercial use.

Re: Infinite Context LLMs: Going Beyond RAG with Extended Minds

#42

Post by selfhoster11 » Wed, Nov 15, 2023, 4:16 PM UTC

mike_hearn wrote: post #40

bsenftner wrote: post #35

Earlier quoted context omitted.

It's open source with a noncommercial use license.

Huh. I wonder if that meets the definition of open source. I didn't notice the special license.

If you go by Open Source Initiative's OSD, then no. Restriction on fields of endeavour is prohibited.

Re: Infinite Context LLMs: Going Beyond RAG with Extended Minds

#43

Post by tayloramurphy » Wed, Nov 15, 2023, 4:18 PM UTC

tj-teej wrote: post #38

Wouldn't this require retraining the model each time your "extended mind" needs more information? One benefit of RAG in my mind is that the "database" can be updated in real-time without needing any model retraining

It's not clear from the paper - but that's what it seems like to me.